Optimize parsing of literal blocks between dollar signs (fixes andialbrecht#491).

andialbrecht · andialbrecht · commit 5a3164d8db9c · 2019-06-26T12:35:02.000+02:00
diff --git a/CHANGELOG b/CHANGELOG
@@ -11,6 +11,8 @@ Bug Fixes
 * Fix splitting of SQL with multiple statements inside
   parentheses (issue485, pr486 by win39).
 * Correctly identify NULLS FIRST / NULLS LAST as keywords (issue487).
+* Fix splitting of SQL statements that contain dollar signs in
+  identifiers (issue491).
 
 
 Release 0.3.0 (Mar 11, 2019)
diff --git a/sqlparse/keywords.py b/sqlparse/keywords.py
@@ -38,7 +38,7 @@ def is_keyword(value):
 
         (r"`(``|[^`])*`", tokens.Name),
         (r"&acute;(&acute;&acute;|[^&acute;])*&acute;", tokens.Name),
-        (r'(\$(?:[_A-Z&Agrave;-&Uuml;]\w*)?\$)[\s\S]*?\1', tokens.Literal),
+        (r'((?<!\S)\$(?:[_A-Z&Agrave;-&Uuml;]\w*)?\$)[\s\S]*?\1', tokens.Literal),
 
         (r'\?', tokens.Name.Placeholder),
         (r'%(\(\w+\))?s', tokens.Name.Placeholder),

-Original file line number
+Diff line change
         (r"`(``|[^`])*`", tokens.Name),
         (r"´(´´|[^´])*´", tokens.Name),
 -        (r'(\$(?:[_A-ZÀ-Ü]\w*)?\$)[\s\S]*?\1', tokens.Literal),
 +        (r'((?<!\S)\$(?:[_A-ZÀ-Ü]\w*)?\$)[\s\S]*?\1', tokens.Literal),
         (r'\?', tokens.Name.Placeholder),
         (r'%(\(\w+\))?s', tokens.Name.Placeholder),